The Age UK almanac of disease profiles in later life A reference on the frequency of major diseases, conditions and syndromes affecting older people in England David Melzer, Joao Correa Delgado, Rachel Winder, Jane Masoli, Suzanne Richards, Alessandro Ble University of Exeter Medical School Ageing Research Group © 2015 UEMS Ageing Research Group, University of Exeter. All rights reserved. Funding This study was supported mainly by Age UK (registered charity number 1128267). JD and AB were supported by the National Institute for Health Research School for Public Health Research (NIHR SPHR). WH was supported by the NIHR Collaboration for Leadership in Applied Health Research and Care for the South West Peninsula. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the NIHR, Age UK or the Department of Health. The team hold a licence to analyse CPRD data and this work was carried out under approved protocol 12_017A4 (June 2015). Acknowledgements This work was supported in part by the National Institute for Health Research School for Public Health Research (NIHR SPHR) Ageing Well programme. SPHR is a partnership between the Universities of Sheffield, Bristol and Cambridge; University College London; The London School for Hygiene and Tropical Medicine; the University of Exeter Medical School; the LiLaC collaboration between the Universities of Liverpool and Lancaster, and Fuse - The Centre for Translational Research in Public Health, a collaboration between Newcastle, Durham, Northumbria, Sunderland and Teesside Universities. The NIHR Collaboration for Leadership in Applied Health Research and Care for the South West Peninsula supported this project to obtain access to the CPRD database. i Contents Introduction ................................................................................. 3 Method .................................................................................................................................................... 4 Prevalence charts ....................................................................... 7 Coronary heart disease ........................................................................................................................... 7 Heart failure ............................................................................................................................................ 8 Atrial fibrillation ...................................................................................................................................... 9 Hypertension ......................................................................................................................................... 10 Stroke and Transient Ischaemic Attack ................................................................................................. 11 Diabetes mellitus................................................................................................................................... 12 Chronic obstructive pulmonary (lung) disease ..................................................................................... 13 Asthma .................................................................................................................................................. 14 Chronic kidney disease .......................................................................................................................... 15 Hypothyroidism ..................................................................................................................................... 16 Recent Cancer ....................................................................................................................................... 17 Epilepsy ................................................................................................................................................. 18 Depression ............................................................................................................................................ 19 Severe mental health conditions .......................................................................................................... 20 Dementia ............................................................................................................................................... 21 Osteoarthritis ........................................................................................................................................ 22 Osteoporosis ......................................................................................................................................... 23 Anaemia ................................................................................................................................................ 24 Falls ....................................................................................................................................................... 25 Fractures ............................................................................................................................................... 26 Urinary and faecal incontinence ........................................................................................................... 27 Skin ulcers and pressure sores .............................................................................................................. 28 Multi-morbidity .......................................................................... 29 Number of co-morbidities by major disease status .............................................................................. 30 Specific co-morbidities with selected major conditions ....................................................................... 31 Appendices ............................................................................... 33 Appendix 1: Prevalence (%) estimates by GP recorded disease in English general practice and hospital records in 2014, with 95% confidence intervals ................................................................................... 33 Appendix 2: Prevalence (%) estimates of GP recorded additional common conditions and syndromes in 2014, with 95% confidence intervals ................................................................................................ 38 Appendix 3: Co-morbidity reference tables .......................................................................................... 40 Appendix 4: Specific co-morbidities with selected major diseases, conditions and syndromes .......... 41 Appendix 5: Electronic record Read codes used in the analyses of geriatric syndromes ..................... 44 References ................................................................................ 61 ii Introduction How many 80 year olds have had a stroke? What proportion of 95 year old men have diabetes? How many older people have seen a general practitioner for problems with incontinence? These types of questions are often asked by patients, carers, doctors and service managers, but, until now, there have been no reliable estimates of the prevalence of common diseases, conditions and syndromes for the oldest groups of people in England. We have therefore obtained and analysed anonymised medical records data on over 600,000 older people – aged 60 and above – from a research database provided by the Government’s health research institute and medicines regulator. The resulting prevalence estimates are presented in this Almanac as graphs, with supporting information to help interpretation (data tables are provided in the Appendices). This compilation is linked to an analysis of diagnostic and treatment trends that we published in Age & Ageing, entitled “Much more medicine for the oldest old” (Melzer et al., 2014), which showed that there was a major increase in recording of disease and intensity of treatment for older people during the last decade, especially for the oldest old. Here, we complement that analysis with the up-to-date estimated of diagnosed diseases. Although older people, especially the oldest old (85 years and over), often have health and social care needs, official statistics and health surveys generally provide patchy information about this group (Sheppard et al., 2012). This is partly because some older people are difficult to reach by traditional surveys, for example those living with frailty or dementia. However, in the UK, general practitioners (GPs) are responsible for the care of the whole population, including those in residential or nursing homes. The availability of anonymised data from GP electronic clinical records, linked to hospital records, makes it possible for the first time to produce estimates of the prevalence of diseases, common conditions and syndromes which are representative of the older population as a whole (meaning those who visit GP practices and/or are attended by GPs). By using a large anonymised sample of records from participating GP practices (see details of database under ‘Methods’ below), the Ageing Research Group from the University of Exeter Medical School have taken a ‘snapshot view’ of the health of the older population across England in 2014. The resulting figures presented in this Almanac provide estimates of: 1. the prevalence of common diseases affecting older people; 2. the prevalence of selected additional common conditions and syndromes, the latter including, for example, incontinence and skin ulcers; and 3. multi-morbidity, providing details of the numbers of diseases that occur together. Despite the many needs of the oldest old, there are no previous studies that we can compare our results to directly. Local studies using groups of older volunteers – e.g. the Newcastle 85+ Study, in which volunteers were aged exactly 85 at baseline (Collerton et al., 2009) and the Medical Research Council Cognitive Function and Ageing Study (CFAS) (Matthews et al., 2013) – provide some overlaps. However, there are no comparable data in the oldest old for the whole of England free of the biases, such as responder bias and loss of volunteers to follow-up, that can severely distort data on older people (Kelfve et al., 2013, Andersson et al., 2012). It should be noted that GP diagnosis and recording of disease in the coded electronic records may not be complete. For example, researchers have reported evidence of under-diagnosis in general practice for conditions including dementia (Connolly et al., 2011), diabetes (Holman et al., 2011) and hypertension (Banerjee et al., 2011). For this 3 reason, we have supplemented the GP-derived data with the hospital admission records from the same patients, thus greatly enhancing the completeness of our estimates. It should be noted that the medical terms recorded by GPs can be complex and sometimes difficult to interpret with certainty. While we have made every effort to include the appropriate codes for each estimate, opinions can differ on details and small differences in coding can influence the reported prevalence. Method The methods used in these analyses were the same as described in “Much more medicine for the oldest old” (Melzer et al., 2014). We used the Clinical Practice Research Datalink (CPRD), which is jointly funded by the NHS National Institute for Health Research (NIHR) and the Medicines and Healthcare Products Regulatory Agency (MHRA). It is a service that makes NHS observational data available for public health research, and has done so since 1987 (http://www.cprd.com). CPRD services are designed to maximise the way anonymised NHS clinical data can be linked to enable many types of observational research and deliver research outputs that are beneficial to improving and safeguarding public health. CPRD is now widely used and its usage has given rise to over 1,500 clinical reviews and papers. CPRD contains the anonymised clinical records of UK patients as entered using diagnostic, symptom and prescription (Read) codes by primary care practitioners. A major advantage is that CPRD includes patients in residential and nursing homes, with essentially complete inclusion of people who have frailty and dependency. The quality of the data is checked by CPRD and it is clear that this is a reliable way to collect medical data on a large scale. We have utilised a complete CPRD dataset for all patients born before 1954 registered with one of the participating general practices in England that take part in the record linkage scheme. We have also accessed linked Hospital Episode Statistics (HES) for the same patients. This dataset collects the diagnoses for each patient admitted to hospital since 1997, and thus provides a powerful addition to the GP records alone. The population included in this dataset is generally representative of the English population in terms of age and sex, when compared with the population projections for England in 2014, developed by Office of National Statistics in 2013. Table 1 describes the population structure. Table 1: Total number of patients (by age and sex)* alive in the Linked CPRD dataset in 2014, meeting eligibility criteria for analysis† Age group Male Female Total N (%) N (%) N (%) 60-64 66,675 (24.1) 67,650 (20.9) 134325 (22.4) 65-69 67855 (24.4) 71021 (22) 138876 (23.2) 70-74 50225 (18.2) 54466 (16.9) 104691 (17.5) 75-79 39073 (14.1) 45584 (14.2) 84657 (14.1) 80-84 28295 (10.2) 37232 (11.6) 65527 (11) 85-89 16143 (5.8) 26290 (8.2) 42433 (7.1) 90-94 6836 (2.5) 15107 (4.7) 21943 (1.7) 95-99 1264 (0.5) 3944 (1.2) 5208 (0.9) 100 150 (0.1) 821 (0.3) 971 (0.2) 4 Total 276,516 (100) 322,115 (100) 598,631 (100) * Figures (N) represent population. † Criteria for eligibility: Registered with the practice for the year of 2014, practice data quality is up to standard. Patients were censored at the earliest date of transfer out of the practice, last collection from the practice or death (data taken CPRD GOLD from the snapshot: November 2014). We studied 15 common conditions (Table 2) that have been used in other studies of the Quality and Outcomes Framework (QOF) (Salisbury et al., 2011, Barnett et al., 2012), a system for monitoring GP practices, but we did not include learning disability or obesity. For defining the medical (Read) codes needed to consider whether a diagnosis was present, we used QOF business rules Version 18.0, October 2010 (Primary care commissioning, accessed April 2012). Diagnoses were identified in general practice coded patient records available in the CPRD Gold dataset, plus the linked hospitalisation records, available from HES for most diseases. We classified each disease as present if the necessary codes appeared in the patients’ records at any time, unless otherwise stated. For example, for cancer and for additional conditions and syndromes, we considered that recent diagnoses were more important (see definitions on each chart). In addition, for a small number of patients for whom GPs had coded that the disease had resolved, diagnoses were not counted. Table 2: Diseases and common conditions and syndromes studied Cardiovascular diseases Neuropsychiatric Hypertension Dementia Atrial Fibrillation Depression Coronary heart disease Epilepsy Heart failure Mental health (Psychoses, Stroke schizophrenia, bipolar affective disorder) Respiratory Endocrine Asthma Diabetes Chronic obstructive pulmonary Hypothyroidism disease Chronic kidney disease Cancer in the previous 5 years (stages 3 to 5) (excluding non-melanoma skin cancer) Additional common conditions Additional syndromes Anaemia Falls Osteoarthritis Fragility fractures Osteoporosis Incontinence (urinary and faecal) Skin ulcers (including pressure sores) There is no real consensus over the key conditions associated with older age (Strandberg et al., 2013), although it is clear that they are linked to disability, frailty, dependence and shorter survival. In addition to the fifteen QOF diseases, we searched the CPRD database for three additional ‘common’ geriatric conditions and four geriatric syndromes associated with older age (Table 2). We considered their inclusion important as, in this age group, common and geriatric conditions have been estimated to be at least as prevalent as other chronic disorders (Cigolle et al., 2007). 5 These geriatric conditions and syndromes are not indicated in QOF, and thus, in order to search for them within CPRD, it was necessary to generate new search terms. The medical literature was examined by two clinicians working independently (or ‘blinded’) of each other, with a third clinical reviewer arbitrating disagreements. The conditions and syndromes were coded as present if a relevant Read code (searched under the categories of ‘symptoms’ and ‘diagnosis’) appeared in the records up to five years (fifteen years for osteoarthritis and osteoporosis) before the beginning of the analysis year (i.e. 2014), to exclude historical diagnoses with no recent mention. Hospital episode statistics (HES) were not used to estimate diagnostic prevalence for the conditions and syndromes as we wanted to focus on longer term disorder rather than acute and possibly short term episodes that might have resolved before patients left hospital. Prevalence graphs The graphs in the following pages present the prevalence (%) of patients with the specified disease, condition or syndrome who were registered with a general practice in England between 1st January 2014 and 31st December 2014. Only data for the periods during which patients were ‘actively’ registered with practices were included in analyses (i.e. we used data from current registration date up to the date of last data collection, transfer out of the practice or death). A small number of apparently ‘non-active’ patients (i.e. those with no clinical or therapy records for the previous three years) were also excluded. ______________________________________ 6 Prevalence charts Coronary heart disease Figure 1: Prevalence of coronary heart disease in English general practice and hospital records in 2014. Coronary heart disease 45 M 40 F 35 30 25 % 20 15 10 5 0 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100+ Age Melzer D et al 2015 Notes: Prevalence of disease is estimated from CPRD records based on clinical codes entered in anonymised GP and hospital records at any time in the patient’s history (with 95% confidence intervals). Coronary heart disease (CHD) is caused by a blockage or interruption to the heart's blood supply, most commonly due to a build-up of fatty substances in the coronary arteries – a condition known as atherosclerosis. Clinically, CHD can manifest as angina or a heart attack. Although still the biggest killer in the UK, mortality rates have decreased by more than 60% since 1968 in most age groups, including those aged 65 to 74 years (Scarborough et al., 2010); there is a similar trend across Europe (Nichols et al., 2013). However, despite mortality improvements, the prevalence of CHD remains high. Figure 1 shows the known higher prevalence of diagnosed CHD in men than women across all older age groups, with over 37% of men aged 85 to 89 years recorded as having CHD. This figure reduces somewhat in ages above 95; however, it remains above 30%. In women, the prevalence of CHD increases with age reaching its highest prevalence (24.9%) in the 100+ year old group. Comparative statistics from the Health Survey for England in 2006 (a community volunteer study) showed self-reported diagnosis (which had been confirmed by their doctor) of CHD in England as 29% for men and 19% for women aged 75 and over (Health Survey for England, 2006). Reducing risk factors, prinicipally smoking (Office of National Statistics, 2011), coupled with more successful interventions and targeted medication, have played a major part in the reducton in CHD mortality, and also in reducing the impact and severity of the disease. 7 Heart failure Figure 2: Prevalence of heart failure diagnoses in English general practice and hospital records in 2014 Heart failure 35 M 30 F 25 20 % 15 10 5 0 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100+ Age Melzer D et al 2015 Heart failure is a serious condition caused by the heart failing to pump enough blood around the body at the right pressure. Breathlessness, tiredness and ankle swelling are the main symptoms (Moser et al., 2014). However, all of these symptoms can have other causes, only some of which are also serious. As is clear from Figure 2, heart failure is common in the older population and increases progressively with advancing age. In addition, co-morbid conditions are increasingly being seen alongside heart failure (Curtis et al., 2008) and there is evidence that the presence of other diseases influences how heart failure progresses (Lam et al., 2011) – see ‘Multi-morbidity’ section from page 29 for details of disease combinations. It is a major public health problem in older people. A definitive diagnosis of heart failure can be difficult as symptoms can be atypical in the older population and hidden by the co-morbidities of respiratory disorders, obesity and venous insufficiency (Manzano et al., 2012, Cleland et al., 2011). The older people included in Figure 2 were considered to have heart failure if at least one recognised diagnostic code was recorded in their GP or hospital discharge records; however, it is possible that the graph underestimates the prevalence of heart failure in the community since under-diagnosis in older and frail people is common (Hancock et al., 2013). The Newcastle 85+ population-based longitudinal study estimated the prevalence of left ventricular heart failure in a community volunteer sample of people aged 87-89 years (including those in institutions and/or cognitively impaired) recruited in 2006-07 (Yousaf et al., 2012). Of the 376 patients in this age group in whom heart function was estimated, half had left ventricular systolic dysfunction or isolated moderate or severe diastolic dysfunction, with almost two thirds of these experiencing difficulty breathing that limited their activities; four fifths of those with significant symptoms of left ventricular dysfunction were undiagnosed (Yousaf et al., 2012). 8 Atrial fibrillation Figure 3: Prevalence of atrial fibrillation diagnoses in English general practice and hospital records in 2014. Atrial fibrillation 35 M 30 F 25 20 % 15 10 5 0 60-64 65-69 70-74 75-79 80-84 85-89 90-94 95-99 100+ Age Melzer D et al 2015 Atrial fibrillation (AF) is a heart condition involving an irregular heart beat with symptoms that can include dizziness, tiredness, shortness of breath and palpitations. AF is known to contribute to strokes and cardiovascular-related mortality, and has a significant impact on quality of life (Valderrama et al., 2005). Increasing with older age and more common in men than women, it has been described as an epidemic in older patients and an important cause of hospitalisation (Steinberg, 2004). In addition, a diagnosis of AF should trigger a patient-doctor discussion about whether the patient should start on warfarin or other anticoagulant drugs to reduce the risk of stroke (NICE, 2014). In our cohort of 598,631 eligible patients registered before 2014, we found that AF was a common diagnosis right into very old age, with 30.8% of men and 25.4% of women aged 95 to 99 years having a recognised diagnostic code for AF at some point in previous GP and hospital discharge records. Men were more likely to have a diagnosis of AF, except in the centenarian group. In a study of 85-year-old volunteers in Newcastle (recruited 2006-07), 14% were found to have atrial fibrillation through 12 lead ECG but a little over a quarter (28%) of these had not been diagnosed in general practice records (Collerton et al., 2009). The prevalence of AF is predicted to continue to increase because of improved survival of people with coronary heart disease, the rising prevalence of diabetes and the growth in the ageing population (Valderrama et al., 2005, Tsang et al., 2005). 9
Description: